INDEX

Explanations

references to military or police operations

oai_token-act-pair · gpt-3.5-turbo

references to "raids" or related concepts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.9.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 warmed

-0.68

milo

-0.64

chell

-0.63

DonaldTrump

-0.63

 Leilan

-0.62

assetsadobe

-0.60

ogyn

-0.60

Ian

-0.59

 Hurricanes

-0.58

issues

-0.56

POSITIVE LOGITS

 raid

0.95

 raids

0.90

 raided

0.84

ers

0.84

ishly

0.82

nard

0.81

 raiding

0.76

iversary

0.75

artment

0.75

 netted

0.73

Activations Density 0.038%

references to military or police operations

references to "raids" or related concepts

No Comments

No Known Activations